Authors' reply

We agree with Schoepf & Heun about the importance of control-sample characteristics in case-control comparisons and the need to consider the possibility of biased or unrepresentative findings. That is why, in our study, we used age as a covariate to take account of the younger age of controls compared with our mood disorder cases. We also discussed the potential limitations on generalisability caused by using a voluntary sample from a single ethnic group.

Schoepf & Heun describe a different methodology (with its own strengths and limitations) to examine a different issue, namely the influence of comorbidity on mortality. It is encouraging that the findings from the two different approaches are consistent in highlighting the clinical importance of comorbidity of physical illness with bipolar disorder. Psychiatrists and mental health services must ensure that patients with bipolar disorder have access to appropriate, high-quality physical, as well as psychiatric, care.